Environ. Sci. Technol. 1904, 18, 694-699
Degradation of Ammonium Nitrate Propellants in Aqueous and Soil Systems Davld L. Kaplan," Patrlcla A. Riley, Davld J. Emerson, and Arthur M. Kaplan US. Army Natick R&D Center, Science and Advanced Technology Laboratory, Natick, Massachusetts 0 1760
The biodegradability of four ammonium nitrate propellants, trimethylammonium nitrate, isopropylammonium nitrate, triethanolammonium nitrate, and hydroxylammonium nitrate, was assessed. The first three compounds were decomposed under aerobic conditions in batch and continuous cultures in a variety of media. Under anaerobic conditions in batch systems and denitrification conditions in continuous flow systems the trimethylammonium nitrate and triethanolammonium nitrate were decomposed while isopropylammonium nitrate was incompletely degraded. No significant buildup of intermediates was observed under any of the conditions studied. All three compounds were readily decomposed in soils at a variety of concentrations. Hydroxylammonium nitrate was found to be chemically unstable at and above slightly acidic pH. Au four compounds tested negative in the Ames screening test for mutagenicity. The contamination of stock solutions of trimethylammonium nitrate and triethanolammonium nitrate with nitrosamines may present the biggest hurdle to overcome in looking toward successful biological treatment of these propellants.
Introduction Substituted ammonium nitrates are under study for use as liquid monopropellants. The potential for biological treatment of these propellents in wastewaters from production, loading, assembly, and packing operations must be assessed as well as the impact on aqueous and soil environments. The purpose of this investigation was to determine the biodegradability of propellants in aqueous and soil systems under a variety of conditions. The identification of intermediates formed during the biotransformation of these compounds and screening of the parent compounds for mutagenicity were also parts of this effort. Experimental Section Chemicals. Trimethylammonium nitrate (TMAN) (methyl-14C, 1.5 mCi/mmol, 98% pure), isopropylammonium nitrate (IPAN) (i~opropyl-l,3-~~C, 5 mCi/ mmol, 98% pure), and triethanolammonium nitrate (TEAN) (ethanol-1,2-14C,5 mCi/mmol, 95% pure) were purchased from California Bionuclear Corp., Sun Valley, CA. Hydroxylammonium nitrate (HAN) (lot R149/ 151), TMAN (lot 128), IPAN (lot 85), and TEAN (lot 190) were provided by the US. Army Ballistics Research Laboratory, Environmental Technology Division, Aberdeen Proving Ground, MD, as 50 wt. % solutions in water. Dimethylamine, methylamine, diethanolamine (DElA), and ethanolamine (ElA) were purchased from Eastman Kodak, Rochester, NY. N-Nitrosodimethylamine (NDMA) was purchased from Aldrich Chemical Co., Milwaukee, WI. N-Nitrosodiethanolamine (NDElA) was purchased from Columbia Organic Chemical Co., Columbia, SC. Batch Systems. Batch studies were conducted with 1000 ppm of TMAN, TEAN, or IPAN in basal salts medium (per liter: K2HP04,1.0 g; KH2P04,1.0 g; MgSO47Hz0, 0.2 g; CaCl,, 0.01 g; NaC1,O.Ol g) with the ammonium nitrates serving as the sole source of carbon and nitrogen. Anaerobic systems (0.025% sodium sulfide) were 694
Environ. Sci. Technol., Vol. 18, No. 9, 1984
incubated unaerated at room temperature while aerobic incubations were flushed with oxygen daily and incubated at 25 "C on a rotary shaker. Active systems under aerobic and anaerobic conditions were run in triplicate with one each of the corresponding sterile controls. Flasks contained 0.34 pCi of either 14C-labeledTMAN or 14C-labeledIPAN or 1.37 pCi of TEAN. Volatile products were collected in 1N NaOH and 1N HC1 traps, and 100-pL media samples were analyzed daily for total radioactivity. Headspace gases not trapped by the acid or base were analyzed by gas chromatography (GC) as described below. At the termination of these experiments the flask contents were centrifuged and filtered, and the resulting cell pellets, filter pads, and filtered broth were counted according to previously described procedures (1). A 50/50 (v/v) dimethyl sulfoxide/acetone mix, 25 mL per flask, was used to extract residual contents on flask walls, and 1mL of this extract was counted. Percent radioactivity within cells harvested from these systems was determined by sonicating the cells for 1 min at 50 W using a Branson sonifier cell disrupter (Danbury, CN), Model W185, with a 0.32-cm tapered microtip. Continuous Systems. Continuous flow systems were run either in BioFlo Model C30 bench top chemostats (New Brunswick Scientific, New Brunswick, NJ) or in modified 500-mL Erlenmeyer flasks. The BioFlo systems used 1500-mL reaction vessels and were run under anaerobic (medium slowly stirred) denitrification or aerobic (stirring and forced aeration) conditions. The modified Erlenmeyer flasks were fitted with a 24/40 ground glass joint, an overflow tube, and a 35 cm long glass tube, 5 mm id., which was suspended in the reaction vessel by a Teflon adapter to deliver nutrient solution to the bottom of the reaction vessel. These vessels were run only under anaerobic denitrification conditions. Nutrient solution was delivered to the reaction vessel by a Rainin Rabbit peristaltic pump (Woburn, MA), and effluent was collected from the overflow line. Retention times were determined by dividing the total volume of medium in the culture vessel by the daily flow rate in milliliters per minute. The continuous flow systems were operated with either TMAN, TEAN, or IPAN as sole or additional carbon and nitrogen sources. Solutions of TMAN were sterilized by passage through a 0.45-pm pore size membrane filter prior to addition to sterilized nutrient reservoirs to avoid losses due to volatilization. TEAN and IPAN were added to the nutrient reservoirs prior to sterilization. All systems were run at room temperature and contained the following quantity of trace salts per liter of filtered lake water: MgS04.7H20, 500 mg; NaC1, 50 mg; CaCl,, 15 mg; FeC13-6H20,10 mg; CuS04.5H20and NaMo04.2H20,10 and 1 mg. In addition to the trace salts the basal salt composition for aerobic systems contained the following per liter: NH4H2P04,2.0 g; KZHPO4,1.0 g; KH2P04,1.0 g; MgS0,.7H20; 0.2 g; CaCl,, 0.1 g; NaC1, 0.1 g, unless otherwise indicated. The basal salt composition for anaerobic denitrification systems contained the following per liter: KN03, 2.05 g; K2P04,0.87 g; trace salt, 505 mg; methanol, 1.4 mL, unless otherwise indicated. Flow rates, pH, oxidation reduction potential, nitrate concentration, methanol concentration, and the concentrations of the ammonium nitrates and their potential
Not subject to U S . Copyright, Published 1984 by the American Chemical Society
decomposition products were monitored weekly or more frequently. Samples, 5-10 mL of influent and effluent from the continuous systems and 2 mL from the batch systems, were taken 1-4 times weekly for analysis by GC or high-performance liquid chromatography (HPLC). Samples were centrifuged at 12OOO rpm for 4 min, and the supernatant was passed through a Swinny stainless steel syringe-type filter holder containing a 0.45-pm pore size membrane filter. All active systems (batch aqueous, continuous, and soil) were inoculated with a mixture of microorganisms from aerobic and anaerobic sewage sludge and soil, which were isolated.in 0.85% KCl and fiitered through Whatman fiiter paper. Stability of HAN. The stability of HAN was assessed over a range of hydrogen ion concentrations. Buffered solutions containing 50 ppm of HAN were adjusted over a pH range of 1.0-6.9. Aliquots of these solutions were assayed spectrophotometrically over 28 days to determine residual concentrations of HAN. Spectrophotometric assays at 705 nm were run on a Perkin-Elmer Lambda 3 spectrophotometer (2). Soil Incubations. A study was carried out to determine the rates of decomposition of the ammonium nitrates at various concentrations in garden soil containing 6.7 % organic matter by ignition and a pH of 5.5. The soil was incubated in 125-mL Erlenmeyer flasks closed with a gas-tight trapping arrangement consisting of an adapter with a vial containing 1 mL of 1 N NaOH. Each flask contained 24.3 g oven dry weight (odw) of sieved (3.35-mm pore size) soil and sufficient moisture to bring the soil up to field capacity. The sterile control flasks were autoclaved for 30 min on three consecutive days. Flasks contained 14C-labeledIPAN, TEAN, or TMAN (0.23,0.27, and 0.14 pCi per flask, respectively) diluted with the corresponding nonradioactive ammonium nitrate to give a final concentration of 50, 500, and 5000 ppm for each of the three compounds. The ammonium nitrate propellants were added as filter-sterilized solutions. Each system was run in duplicate, and the traps were changed 3 times the first week and weekly or every 2 weeks thereafter for the remainder of the experiment. Incubations ran for a total of 103 days. The second soil experiment was run for a total of 7 months in 125-mL Erlenmeyer flasks. These flasks were modified with two trapping tubes to allow for continuous collection of gaseous products in 1.0 mL of 1N NaOH and separately in 1.0 mL of 1N HC1. The flasks contained one of the following three matrices: (1) 29 g odw of garden soil (pH 6.4,42.5% organic matter by ignition, passed through a 3.35-mm pore size sieve) and 15 mL of distilled water, (2) the same soil but flooded with 55 mL of distilled water, or (3) sand, 70 g odw (Fischer Scientific Co., Fair Lawn, NJ, washed and ignited, pH 7.9,0% organic matter), and 15 mL of distilled water. Sterile flasks were autoclaved as described, and a filter-sterilizedsolution containing both the 14C-labeled and the unlabeled ammonium nitrate compound was added after autoclaving. Incubation flasks contained IPAN, 0.32 pCi per flask, TMAN, 0.17 pCi per flask, or TEAN, 0.024 pCi per flask, diluted to 100-ppm final concentration with the corresponding unlabeled ammonium nitrate. For the most part, acid and base traps were changed 4 times the first week of incubation, 2 or 3 times a week for the next 3 weeks, and once or twice per month for the duration of the experiment. Flasks were incubated a t room temperature throughout the experiment. Final pH and oxidation-reduction measurements were made on a Corning Model 130 pH meter with soil
samples diluted with an equal volume of degassed, distilled, deionized water when necessary. Thin-Layer Chromatography. Thin-layer chromatographic (TLC) analysis of TMAN, IPAN, and TEAN was run in 1-butanol/acetic acid/water (4/ 1/51 on cellulose plates. Spots were visualized with iodine vapors. The R values were 0.42,0.47, and 0.58 for TEAN, TMAN, and IPAN, respectively. TEAN was resolved from DElA and E1A in a solvent system consisting of chloroform/methanol/ammonium hydroxide (47/47/6) on silica plates and resulted in R, values of 0.40,0.50, and 0.80 for ElA, DElA, and TEAN, respectively. Spots were visualized with iodine vapors. Several other reagents were examined for visualization of the three compounds (cobalt(I1) thiocyanate, potassium ferricyanide-ferric chloride, ninhydrin, and ninhydrincupric nitrate). However, only ethanolamine could be visualized with several of these reagents. Gas Chromatography. GC analysis of TMAN, TEAN, and IPAN as the free amines and DMA, MA, methanol, DElA, and E1A was accomplished on a Hewlett-Packard Model 5840A GC with a flame ionization detector (FID). A 3-m glass column packed with 10% Carbowax 20M and 2% KOH on 8O/lOO Chromosorb WAW and conditioned at 170 "C with 45 mL/min nitrogen carrier flow for 24-48 h was used for analysis of TMAN, DMA, MA, IPAN, and methanol. Details on the development of this method and conditioning requirements were previously described(3). Operating conditions were with injector, oven, and FID temperatures of 150, 30, and 230 OC, respectively. All sample injections were 2 pL, and injections of water were used to keep the column properly conditioned. GC separation of TEAN, ElA, and DElA was accomplished with Tenax-GC packing. TEAN and DElA were determined with an oven temperature of 260 "C while E1A was analyzed at an oven temperature of 180 "C. For all three compounds the injector temperature was 280 OC, the FID temperature was 380 OC, and carrier gas flowed at 60 mL/min. The use of Tenax-GC for analysis of ethanolamines has been previously described (4). The detection limits, as calculated by the method of Hubaux and Vos (5) were as follows: TMAN, 11.4 ng; DMA, 29.6 ng; MA, 32.8 ng; IPAN, 15.4 ng; methanol, 6.0 ng; TEAN, 175.8 ng; DElA, 201.4 ng; ElA, 157.0 ng. GC analyses for headspace gases for the batch experiment with TEAN, TMAN, and IPAN incubated as the sole sources of carbon and nitrogen were determined on a Hewlett-Packard Model 5880 GC with a thermal conductivity detector. The column was a 2.67 m long by 0.32 cm diameter stainless steel column packed with Carbosieve S 120/140 mesh purchased from Supelco, Inc. (Bellefonte, PA). The oven temperature was programmed from 35 to 250 "C at 15 OC/min, and injector and detector temperatures were 150 and 275 OC, respectively. Helium carrier gas flowed at 39 mL/min, and injections were 0.5 mL. High-Performance Liquid Chromatography. In order to determine whether nitrosamines might form during the biological decomposition of these ammonium nitrate propellanb under denitrification conditions, HPLC was used to quantitate NDMA and NDElA in influent and effluent samples from continuous flow systems. Analyses were performed on a Waters Associates (Milford, MA) HPLC equipped with two Model 6000A solvent delivery pumps, a Model 441 detector set at 254 nm, a Model 720 system controller, and a Model 730 data module. All nitrosamine analyses were run on a pBondapak C18 reverse-phase stainless steel column, 3.9 mm X 30 cm (Waters Associates). Envlron. Scl. Technol., Vol. 18, No. 9, 1984
695
Table I. Chronology of Changes and Results of Analysis for the Anaerobic Denitrification Continuous Flow System with TMAN (100 ppm) time, days
retention time: days, f f 1 SD
media composition
PH, fflSD
denitrification, %, f f 1 SD
0-55 56-74 75-91 92-98 99-130
4.6 f 0.8 2.1 f 0.5 1.6 f 0.5 5.5 f 1.2 4.7 A 0.9
basal saltC d d d basal salts'
8.6 f 0.2 8.4 f 0.2 8.4 f 1.0 8.5 f 0.0 9.0 f 0.2
99.0 0.0 74.3 f 3.0 57.0 f 24.0 93.0 0.0 83.2 f 10.8
degradation,b %, f f 1 SD TMAN methanol
*
99.7 f 0.8 99.5 f 0.6 99.5 f 1.1
*
92.4 f 5.4 82.8 f 7.1 60.7 f 3.1 ND NAf
NDe 99.8 f 0.5
"Volume of medium in the culture vessel divided by daily flow rate. bPercent change in concentration between influent and effluent samples; analysis by GC. 'KNOB, 4.93 g, and methanol, 4.1 mL. Ratio of grams of total organic carbon to grams of total nitrogen as nitrate (C/N) = 1.8. dNo change. eNo data. fSodium acetate, 6.96 g/L, instead of methanol as the electron donor. #Not applicable.
Table 11. Chronology of Changes and Results of Analysis for the Anaerobic Denitrification Continuous Flow System with IPAN (100 ppm)' time, days
retention time,b days, f 1 SD
PH, fflSD
denitrification, %,f*lSD
0-64 65-97
4.3 f 0.7 8.7 f 1.5
8.1 f 0.5 8.2 f 0.1
99.0 f O.Od 98.9 f 0.4
*
degradation: %, f f 1 SD IPAN methanol 32.1 f 19.2 42.8 f 5.0
95.6 f 5.9 100.0 f 0.0
Medium was basal salts, C/N = 1.5. Volume of medium in the culture vessel divided by daily flow rate. Percent change in concentration between influent and effluent samples; analysis by GC. dDenitrification was 59.3 f 13.7% for the first 34 days and 99.0 f 0.0% after that oeriod.
NDMA was analyzed with water as mobile phase at a flow rate of 2.5 mL/min. Injection volumes varied up to 200 pL, and the retention time was about 3.5 min. NDElA was analyzed with water containing 0.01 M K2HP04as the mobile phase flowing at 1.5 mL/min through two CI8 pBondapak reverse-phase stainless steel columns (3.9 X 30 cm each) in tandem. Injection volumes varied up to 200 pL, and the retention time was around 7 min. Scintillation Counting. The acid-base traps were counted for radioactivity in a Packard Tri Carb Model 3255 liquid scintillation counter with Aquasol 2 cocktail (New England Nuclear, Boston, MA). Sodium hydroxide traps were treated with Aquasol 2 and distilled water, while acid traps were treated only with the cocktail. Mutagenicity Testing. The Ames screening test for mutagenicity was performed according to standard procedures (6). TEAN, TMAN, IPAN, and HAN were tested with and without metabolic activation at 5, 50, 500, and 5000 pg per plate against five strains of Salmonella typhimurium (TA98, TA100, TA1535, TA1537, and TA1538). The tests were run in triplicate.
Results Batch Studies. The results from the batch experiment with the ammonium nitrate propellants incubated as the sole source of carbon and nitrogen are presented in Figure 1. For each of the three ammonium nitrates the total recoveries of radioactivity in the alkaline traps approached 45% in the aerobic systems and between about 5 and 15% in the anaerobic systems. The lag prior to rapid onset of mineralization was shorter for TEAN and longest for TMAN under the aerobic conditions. The corresponding control flasks in this study are not illustrated because no significant levels (less than 0.1% of the total) of radioactivity were recovered in acid or base traps during the experiment. The change in radioactivity in the culture broth of the active aerobic cultures showed the drop in residual activity as the compounds were degraded. This decrease was earliest for TEAN while all flask contents appeared to be inhibitory to further activity after about 60% had 696
Environ. Sci. Technol., Vol. 18, No. 9, 1984
1'
TMAN
[
IPAN
1
TEAN
1
TIME (DAYS) Figure 1. Decomposition of radioactively labeled TMAN, IPAN, and TEAN under aerobic (open circles) and anaerobic (soikl circles) batch culture conditions as the sole source of carbon and nitrogen.
been lost from the medium. No significant counts were recovered in the acid traps for any of the incubations. fractionation of the media indicated that filter pads and flask residues retained a few percent of the total counts while 7.1% of the counts were found associated with the cell pellet. Continuous Flow System Studies. System changes, measurements of pH, denitrification efficiency, and degradation of the ammonium nitrates and methanol in the continuous flow systems are presented in Tables I-IV. In general, denitrification systems ran efficientlywhen carbon to nitrogen ratios and retention times were suitably adjusted for each of the systems. The pH of the effluents from the denitrification systems were in the alkaline range as expected in a successfully operative denitrification system. The concentrations of the ammonium nitrates and methanol were monitored by GC. Under anaerobic denitrification continuous flow conditions TMAN, 100 ppm, was efficiently degraded (Table I) at 4.6-, 2.1-, and 1.6-day retention times. This was also the case when sodium acetate was substituted for methanol as the electron donor. With methanol in the system, the efficiency of utilization and the denitrification efficiency decreased as the retention time was shortened. Thus, even
Table 111. Chronology of Changes and Results of Analysis for the Anaerobic Dinitrification Continuous Flow System with TEAN' time, days
TEAN concn, ppm
retention time,b days, f f 1 SD
PH, f f l S D
denitrification, %, f f 1 SD
TEAN degradation: %, if 1 SD
0-29 30-76 77-100 101-161
1000
9.3 f 1.1 4.1 f 0.4 9.5 f 1.0 7.6 f 0.9
8.0 f 0.1 7.9 f 0.1 7.6 f 0.0 7.8 f 0.1
56.0 f 0.0 95.3 f 4.7 99.0 f 0.0 94.3 f 4.2
NDd 31.8 f 17.0 60.8 f 20.8 63.3 f 16.5
e e
500
Medium was basal salts, C/N = 1.5. Volume of medium in the culture vessel divided by daily flow rate. tration between influent and effluent samples; analysis by GC. No data. e Unchanged.
Percent change in concen-
Table IV. Chronology of Changes and Results of pH Analysis for the Aerobic Continuous Flow System with TMAN (100 ppm)
days
retention time,' days, f f 1 SD
0-44 45-79 80-93 94-114 115-135 136-177
4.0 f 0.5 2.4 f 1.0 1.0 f 0.1 5.4 f 1.7 2.6 f 0.5 4.2 f 0.3
media composition
PH, fflSD
TMAN degradation,b %,fflSD
nutrient broth (4 g/L)
8.1 f 0.2 8.2 f 0.3 8.1 f 0.0 7.3 f 1.1 6.0 f 0.0 6.7 f 0.1
99.4 f 0.9 100.0 f 35.6 51.5 f 35.6 93.5 f 9.2 99.3 f 1.2 100.0 f 0.0
C
C
basal salts C
basal saltse
"Volume of medium in the culture vessel divided by daily flow rate. bPercent change in concentration between influent and effluent samples; analysis by GC. 'Unchanged. Without NH4H2P04.
Table V. Chronology of Changes and Results of pH Analysis for the Aerobic Continuous Flow System with IPAN (100 ppm)
days
retention time,' days, if 1 SD
media composition
PH, fflSD
IPAN degradation,b %,iflSD
0-25 25-56 57-74 75-95 96-116 117-137
4.1 f 0.4 2.0 f 0.2 1.1f 0.2 4.7 f 0.3 2.8 f 0.3 4.4 f 0.2
nutrient broth (4 g/L)
8.9 f 0.1 8.8 f 0.0 8.6 f 0.1 6.2 f 0.0 6.1 f 0.2 6.5 f 0.4
110.0 f 0.0 86.5 f 13.3 42.7 f 23.8 100.0 f 0.0 94.0 f 4.2 97.0 f 4.2
C
C
basal salts C
basal saltsd
'Volume of medium in the culture vessel divided by daily flow rate. bPercent change in concentration between influent and effluent samdes: analvsis bv GC. 'Unchanged. Without NH,H,PO,.
Table VI. Chronology of Changes and Results of pH Analysis for the Aerobic Continuous Flow System with TEAN
days 0-10 11-48 49-81 82-111 112-147
TEAN concn, ppm
retention time,' days, if 1 SD
media composition
PH, fflSD
TEAN degradation,b %, f f 1 SD
1000
nutrient broth (4 g/L)
C
4.0 f 0.2 8.7 f 1.2 8.3 f 1.9 8.0 f 0.7
8.6 f 0.0 8.9 f 0.1 8.6 f 0.1 6.6 f 0.3
91.5 f 7.8 78.0 f 6.1 77.5 f 17.0 50.3 f 16.7
C
5.8 f 0.3
6.3 f 0.2
58.0 f 31.1
C
500
C C
basal salts + glucose (1 g/L) basal salts
'Volume of medium in the culture vessel divided by daily flow rate. *Percent change in concentration between influent and effluent samples; analysis by GC. Unchanged. I
when the denitrification efficiency was adversely effected by shortened retention times, TMAN was still effectively degraded. IPAN (Table 11) was incompletely degraded in the denitrification system a t a 4.3-day retention time, and there was little improvement at an increased retention time, 8.7 days. Denitrification efficiency and methanol utilization were high throughout the operation of this system. TEAN, 1000 ppm, was not efficiently degraded at a 4.1-day retention time, while higher percentages of decomposition were achieved at longer retention times. (9.5and 7.6-day retention times with 1000 and 500 ppm of TEAN, respectively) (Table 111). In all instances except
for the initial acclimation period, denitrification efficiency was high. TEAN was studied at 500 ppm concentrations because of the higher detection limits by GC analysis. Results from aerobic continuous flow systems are presented in Tables IV-VI. TMAN (Table IV) was readily degraded at 4.0- and 2.4-day retention times, but efficiency decreased at a 1.0-day retention time in nutrient-rich solution. In basal salt medium both with and without supplemental nitrogen TMAN was effectively decomposed. The basal salt system without nitrogen represents a culture medium in which TMAN serves as the sole source of carbon and nitrogen. In Table V in nutrient-rich solution, IPAN was completely degraded at a 4.2-day retention time and proEnviron. Sci. Technol., Vol. 18, No. 9, 1984
897
sand medium and lowest in the soil. Mutagenicity Screening. The results of Ames mutagenicity test of the four ammonium nitrate propellants indicated no evidence for mutagenic activity, as all mutation rates fell within background levels. It was noted, however, that HAN was toxic to S. typhimurium at the higher concentrations tested.
2
4
6
2
4
6
2
4
6
TIME (MONTHS)
Figure 2. Summary data on the decomposition of radioactively hbeled TMAN, IPAN, and TEAN In soil (opencircles), flooded soil (solkl circles), and sand (open squares).
gressively less efficiently degraded as the retention time was shortened. In basal salt medium with and without supplemental nitrogen IPAN was readily degraded. TEAN was effectivelydecomposed at a 4.0-day retention time and somewhat less efficiently degraded as the retention time was lengthened at both 1000 and 500 ppm (Table VI). TEAN was incompletely metabolized when present in a basal salt medium both with and without supplemental glucose. Nitrosamines. Influent and effluent samples from the TMAN continuous flow denitrification system were monitored for NDMA by HPLC. Low ppb levels of NDMA were present in both influent and effluent samples, and for the most part, concentrations were slightly higher in the effluent. Hydroxylammonium Nitrate. HAN was stable at pH 4.9 and below, while at higher pH values it degraded rapidly. Soil Studies. In the first set of soil incubations the effect of concentration of the ammonium nitrates on rates of mineralization in soil was examined. At concentrations of 50,500, and 5000 ppm the total percent degraded was not significantly different for each of the three compounds. At 5000 ppm the initial rates of degradation were somewhat lower for TMAN and IPAN, but this lag was no longer evident after the first month. Only trace amounts of radioactivity were recovered in alkali traps from the control flasks during the study. The total percent degradation after 103 days was between 45 and 60% for all of the active flasks. In the second set of soil incubations the effect of soil conditions on rates and total degradation of the ammonium nitrates over 7 months was examined (Figure 2, data summerized by month, except for the first month). There was rapid initial rate of degradation during the first month and a relatively gradual increment thereafter in all the active systems. At most 4.2% of the total counts were released into the acid traps from the active systems during the 7 months and considerably lower amounts from the sterile controls. Between 0.1% and 2.1% of the total counts were recovered in base traps from sterile controls during the 7 months for the three compounds in the three media. A total release of between 53.7% and 74.5% was found for all the active flasks during the study under the different soil conditions. The initial rates of degradation of TMAN were highest in soil, lower in flooded soil, and lowest in sand; however, by the end of the incubation period the total percentages were within 5%. A similar pattern of initial reaction rate was found in the soil flasks incubated with IPAN. After 7 months there was a higher total release from the flasks containing flooded soil. Initial rates of release of 14C02from TEAN were similar under all three conditions while total release was highest in the 698
Environ. Sci. Technol., Vol. 18, No. 9, 1984
Discussion The results from the biological studies with TEAN, TMAN, and IPAN indicate that all three compounds are biodegradable. Under aerobic conditions they are decomposed, and there is no evidence for the formation of significant concentrations of intermediates. Under anaerobic denitrification conditions TEAN and TMAN were biodegraded, while IPAN was only partially decomposed. Under anaerobic conditions, there is also no indication for significant formation of intermediates. For the most part, intermediates were detectable only if the system was perturbed. Under aerobic batch conditions all three ammonium nitrates are mineralized in a minimal medium in which they are the sole source of carbon and nitrogen. These findings suggest that development of a biological approach to treat process waters containing these compounds would be a feasible one. The system should have a degradative capability over a wide range of environmental conditions. In organic-poor or organic-rich soils, or in flooded soils at concentrations up to 0.5%, the three compounds are readily mineralized. As in most batch systems, the initial rates of decomposition are very rapid with a subsequent leveling off of activity and gradual release of carbon dioxide. This was true for the aqueous and soil batch systems studied. Inhibitory effects may develop in these systems which prevent 100% mineralization. Similarly, incorporation of some of the material into microbial biomass would account for a slow turnover and subsequent release of the 14C label. HAN was found to be chemically unstable above a pH of about 5.9. This indicates greater instability for the nitrate salt of hydroxylamine than the free amine. The free amine was reported to be unstable above pH 6.8 in the work of Frear and Burrell (2). Reports in the literature indicate that hydroxylamine rapidly disappears from soils through a number of chemical reactions with inorganic and organic soil components (7,8). With a demonstrated instability above pH 5.9 and with the numerous potential reactions in soils, it would be expected that HAN would not persist under most environmental conditions. Hyphomicrobium sp, was the predominant microorganism identified in culture vessels from the denitrification systems with methanol as the principal carbon source. Pathways have been established in the literature for the degradation of some amines (9) corresponding to the nitrate salts examined in this study (trimethylamine is metabolized successively to dimethylamine and methylamine). The presence of NDMA in influent as well as effluent samples from the continuous flow systems with TMAN prompted an investigation of the stock propellant solutions. TMAN and TEAN stocks were found to be contamined with NDMA (2.8 ppm) and NDElA (8.4 ppm), respectively. IPAN contained neither of these contaminants. All of these findings were confirmed by analysis on either a GC or HPLC thermoelectron analyzer (Therm0 Electron Corp., Waltham, MA) using either a model 543 or Model 502 nitrosamine specific detector. The presence of these contaminating nitrosamines may present the major difficulty is using a biological approach to alleviate po-
Environ. Sci. Technol. 1004, 18, 699-705
(3) Emerson, D. J.; Kaplan, D. L.; Kaplan, A. M. U.S.Army Natick R&D Labs, Natick, MA, 1982, Technical Report TR-831004. (4) Saha,N. C.; Jain, S. K.; Dua, R. K. Chromatographia 1977, 10,368-371. (5) Hubaux, A.; Vos, G. Anal. Chem. 1970,42,849-855. (6) Ames, B. N.; McCann, J.; Yamasaki, E. Mutat. Res. 1975, 31 , 347-364. (7) Brenner, J. M.; Blackmer, A. M.; Waring, S. A. Soil Boil. Biochem. 1980,12,263-269. (8) Nelson, D. W. Proc. Indiana Acad. Sci. 1978,87,409-413. (9) Harder, W.; Attwood, M. M. Adv. Microb. Physiol. 1978, 17, 303-359.
tential pollution hazards associated with these compounds.
Acknowledgments We thank J. Knapten a t BRL for his helpful communications. We thank J. Pierce and S. Cowburn for their technical assistance. &&;istry NO.TMAN, 25238-43-1; IPAN, 87478-71-5; TEAN, 27096-29-3; HAN, 13465-08-2.
Literature Cited (1) Kaplan, D. L.; Kaplan, A. M. Environ, Sci. Technol. 1982, 16, 566-571. (2) Frear, D. S; Burrell, R. C. Anal. Chem. 1955,27,1664-1665.
Received for review September 26, 1983. Revised manuscript received March 5, 1984. Accepted March 21, 1984.
Study of Rapping Reentrainment Emissions from a Pilot-Scale Electrostatic Precipitator P. Vann Bush Southern Research Institute, Birmingham, Alabama 35255-5305
q/section = 1 - exp(-x/N)
A test program was conducted to determine the quantity and size distribution of rapping reentrainment emissions from a large pilot scale electrostatic precipitator at the TVA Bull Run Steam Plant. The precipitator current density and specific collection area were varied during the test program. The data were compared to information from full-scale precipitators which had been used to derive generic relationships for rapping reentrainment incorporated in the mathematical model of precipitation.
Introduction A characterization of the rapping contribution to emissions from a pilot-scale electrostatic precipitator (ESP) was performed in order to evaluate an hypothesis that rapping-related emissions were responsible for the discrepancy between performance measurements and ESP model predictions of performance. The pilot-scale ESP has a gas volume capacity of 30,000 acfm. The internal arrangement of the system is shown in the schematic sectional side view in Figure 1. The precharger field was not energized for this study. The collector stage consists of four standard Lodge-Cottrell design electrical fields of 9-, 6-, 6-, and 9-ft lengths in the direction of gas flow. There are three rapping fields: fields 2 and 3 share 12 ft long collection plates. The total plate height in all fields is 12 ft, with an active height of 10 f t 3 in. There are 13 gas passages in the collector stage, with 10-in. width, giving a total active collection surface of 7995 ft2. The discharge electrodes installed in the system were 3/a-in. diameter wires mounted in the Lodge-Cottrell mast design frame. The pilot scale ESP system has been more thoroughly described in other reports (1,2). The ESP model used for predicting the performance of the pilot-scale ESP incorporates a rapping loss calculation based on a limited field study of full-scale electrostatic precipitators (3, 4 ) . The study was conducted at four cold-side ESP installations and two hot-side ESP’s. In order to discover a relationship useful in predicting rapping losses, the limited data were plotted as a function of dust calculated to have been removed by the last field of the ESP. The dust removal in the last field was approximated by applying the Deutsch equation in the form 0013-936X/84/0918-0699$01.50/0
(1)
where x = -In (1- vo),vo = overall mass collection fraction determined from mass train measurements, and N = number of ESP sections. These data are plotted in Figure 2. The exponential relationships shown for the hot-side and cold-side ESP data were presented in the rapping study (4) as merely aids for interpolation. It was also stated that additional data under a wider variety of conditions were required to verify the validity of this approach. With these caveats the exponential relationships shown in Figure 2 were placed in the model and have been in use since 1978. The relationships are employed in the ESP model to predict the total mass emissions due to rapping. The information generated in this way is all that is used in this model to produce penetration (or collection efficiency) numbers “corrected for rapping reentrainment”. Additional information about the rapping emissions is required to define the fractional penetration (the penetration of particles as a function of their diameter). To provide this information for the model, the data from the limited study were used to compute the apparent rapping puff size distribution at each plant. Figure 3 shows these data. A representative rapping puff size distribution was extracted from the average of the data shown in Figure 3 and approximated by a log-normal size distribution having a mass median diameter (MMD) of 6 Fm and a standard deviation of 2.5. When the ESP model has determined the quantity of mass in the rapping emissions using the exponential relationships in Figure 2, the mass is divided into size fractions coinciding with the log-normal distribution. With the limitations in the data base from which the ESP model treatment of rapping reentrainment was derived, and the differences which may exist between fullscale and large pilot-scale operation, it is reasonable to expect the model is not accurately predicting the contribution of rapping reentrainment to the emission of the pilot ESP.
Test Program Tests were designed to quantify the percentage of emissions due to rapping in the pilot ESP. The measurements selected to gather this information were method
0 1984 American Chemical Soclety
Environ. Sci. Technol., Vol. 18, No. 9, 1984 6QQ
